Accession Number:

ADA495840

Title:

Predicting Solar Protons: A Statistical Approach

Descriptive Note:

Master's thesis

Corporate Author:

AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OH SCHOOL OF ENGINEERING AND MANAGEMENT/DEPT OF ENGINEERING PHYSICS

Personal Author(s):

• Spaulding, Jonathan C.

Report Date:

2009-03-01

Pagination or Media Count:

94.0

Abstract:

A small fraction of solar flares are accompanied by high energy 10 MeV protons. These events can cause degradation or failure of satellite systems and can be harmful to humans in space or in high altitude flight. For risk management purposes, the Air Force is interested in predicting these events. Several algorithms exist to do this operationally, but none predict when these events will occur with much accuracy. Here, we analyzed 3610 M1 and greater flares including 106 with proton events from the GOES sensors from 1 Jan 1986 to 31 Dec 2004 to produce new results, including a full scale comparison and optimization for all the algorithms. In every case, optimization leads to increased prediction ability. This research also produced a new algorithm based on the Garcia algorithm, which functions better than any other operational algorithm. This model, Garcia 2008, predicts with a skill score of .526, an improvement from .342. This new model is the best at prediction of all models measured.

Descriptors:

• *SOLAR PHYSICS
• *PROTONS
• *STATISTICAL PROCESSES
• OPTIMIZATION
• DETECTORS
• SKILLS
• DEGRADATION
• PREDICTIONS
• MODELS
• HUMANS
• COMPARISON
• FAILURE
• ACCURACY
• HIGH ENERGY
• MATHEMATICAL PREDICTION
• ARTIFICIAL SATELLITES
• HIGH ALTITUDE
• SOLAR FLARES
• AIR FORCE
• RISK MANAGEMENT
• MATHEMATICAL MODELS
• ALGORITHMS

Subject Categories:

• Astrophysics
• Statistics and Probability

Distribution Statement:

APPROVED FOR PUBLIC RELEASE